Re: [PATCH 08 of 10 v3] libxl: enable automatic placement of guests on NUMA nodes

[Ian Campbell <Ian.Campbell@citrix.com>]

On Fri, 2012-07-06 at 14:03 +0100, Dario Faggioli wrote:
> On Fri, 2012-07-06 at 11:55 +0100, Ian Campbell wrote: 
> > > Once we know which ones, among all the possible combinations, represents valid
> > > placement candidates for a domain, use some heuistics for deciding which is the
> > > best. For instance, smaller candidates are considered to be better, both from
> > > the domain's point of view (fewer memory spreading among nodes) and from the
> > > system as a whole point of view (fewer memoy fragmentation).  In case of
> > > candidates of equal sizes (i.e., with the same number of nodes), the amount of
> > > free memory and the number of domain already assigned to their nodes are
> > > considered. Very often, candidates with greater amount of memory are the one
> > > we wants, as this is also good for keeping memory fragmentation under control.
> > > However, if the difference in how much free memory two candidates have, the
> > > number of assigned domains might be what decides which candidate wins.
> > 
> > I can't parse this last sentence. Are there some words missing after
> > "how much free memory two candidates have"?
> > 
> Ok, I see. What about something like the below:
> 
> "However, if the amount of free memory of two candidates is very
> similar, we look at how many domains are assigned to each candidate, and
> take the one that has fewer of them."

Looks good, thanks.

Is it still "very similar" now that it is a total order or is it
actually a strict equality now? (or an equality down to the precision of
a float...)

> > If you want to post the corrected text I think we can fold it in while
> > applying rather than reposting (assuming no other reason to repost crops
> > up).
> > 
> Let's see... I'm fine either way...

I think there's a few other fixes to be made so it'll have to be a
repost.

Ian/.